A nostalgic look at Diversification

Everyone knows what diversification is. From early days as schoolchildren we were taught "you should not put all your eggs in one basket."

How impractical is that advice? I keep chickens. Visualise collecting and carrying eggs in several baskets. You would surely drop one basket in time. A more realistic approach is to carry one basket and take damn care not to drop it. I never have nor have the kids or grandkids!

The same principle applies to shares. The more experienced and knowledgeable about investment you are, the more you can specialise on just a few investments. Rupert Murdoch, the late Kerry Packer and Warren Buffet are all known for highly concentrated investments.

Prudence also dictates that investments are spread sufficiently so that the risk of one unfortunate chance happening does not impact heavily on your wealth. But how much do you need to diversify?

Until 1964 no-one had really tested how many shares should be in a portfolio. In that year an American called Benjamin King published the results of his research which showed that sufficient diversification can be achieved with as few as 10 stocks chosen randomly, provided you invest approximately equal amounts of money in each.

I obtained similar results in 1975 during my post-graduate research in the UK market (Published in 1983 as chapter 5, "The Role of Risk in Industry Analysis, John Aldersley,
in a book published by MacMillans "Modern Portfolio Theory andFinancial Institutions" – edited by Desmond Cornerand David G Mayes, ISBN  0333307755).

An almost identical relationship holds for the Australian market too.

Professional Fund managers usually own considerably larger numbers of securities. Admittedly not as many as when I first started.

I first started in fund management back in 1976, when very few people in senior positions of investment management had any formal investment training. They had generally moved across from banking. They were simply unaware of the research and did not have the skills to wade their way through the heavy mathematics in search of a practical tool. Capital asset pricing theory was still limited to a few academics in the UK.

They felt a safety in investing in large numbers of securities. After all, this was how it had always been done in the past. Then there was the British class distinction between the fund managers, in London, and the administration staff, based half a day away. Not one of the fund managers had ever visited the administration operation, and so there was a disregard of the administration effects and costs of this lack of discipline.

Listed Investment Trust Companies started in the UK as early as the 1870's. Foreign and Colonial was the first. The passing of the years had led some funds to accumulate thousands of individual securities.

I was appointed assistant investment manager for the Charter Trust & Agency in 1976. Charter was one of about eight listed investment companies managed by the Kleinwort, Benson stable.  

As I reviewed the portfolio I counted about 5,300 individual securities from all around the world. Yet the total value of the portfolio was only around 40 million pounds. Many of the securities were worth less than 100 pounds sterling. Some of the otherwise worthless Chinese and Russian bond certificates from the turn of the century were worth several thousand pounds as collector items.

I was not given responsibility for the 4000 or so overseas holdings. These were managed (sic) by two other individuals. I reviewed the UK holdings. As the official engineering analyst, I felt most competent in first reducing this component. I was appalled to find that Charter held 115 UK engineering companies, most of which were heading towards oblivion. Their investments included a stagecoach wheel maker, the RHP ball bearing manufacturer, a cooper (a beer barrel hoop supplier), and several run down foundries. We had the lot.  I sold 112 separate holdings leaving just two process engineering contractors and one Sheffield steel producer. I even sold GKN (the equivalent to Australia's BHP) and was promptly accused of selling Britain short.

I tried to introduce three small growth companies from the electronics sector. Racal (military radio packs), United Scientific (laser range finders and sonar buoys), and Amstrad (consumer electronics). KB had just listed Amstrad. I had bought shares in each of these companies for the BBC Pension Fund. All three recommendations were rejected by the trust directors because the dividend yield was too low. Charter rarely bought anything on less than a 5% yield. The share prices all rocketed upwards in value and have since become major global companies. Racal now owns Vodaphone. Any one of them would have transformed the value of this trust over the past thirty years or so.

I learned some valuable lessons from this period. How could I possibly add value to a portfolio if I did not know what most of the companies did? The trust was supposed to be an active manager, but looking through board papers I got the impression that a few tiny trades were made each month simply to show some activity to justify the monthly management fees. It hardly seemed worth the effort to thoroughly investigate a company, including a site visit to the North of England, and then buy just 50,000 pounds worth in a 40 million pound portfolio.

I also realised that reducing the administration costs by reducing the number of holdings would have a much greater impact on the bottom line (for both clients and the bank) than my individual efforts in stock selection.

I visited the administration operation then based in Newbury. Everything in those days was manually typed in foolscap. Costs were supposedly charged back to the trusts but the operation was never really costed so the investment trust division operated around break-even. This was despite continuous management since the 1870's!

I also saw at first hand how holding too many securities dilutes your winners so the effort is disproportionate to the effect. Doubling your money on 50,000 pounds in three months had an insignificant effect.

The Pension Fund side of the business was growing quickly and in 1979
as a young "gunslinger" I was given prime responsibility for managing the UK holdings of the BBC and Blue Circle portfolios. In those days the BBC Pension fund was split between seven managers. We were the seventh ranking manager and two were to be dropped in six months time. Looking back I realise I was allocated this fund for political reasons.

In contrast to the trust companies the Pension Funds were managed in a disciplined way. Portfolios were generally much larger yet held no more than 50 securities, and while asset allocations varied according to the agreement with the respective trustees, there was a very common approach taken. In fact each manager had to conform more or less with benchmark weightings on a sector basis, leaving some discretion as to the individual stock selections within each sector.

The Wood Mackenzie survey of six months later showed that our BBC portfolio ranked second in performance of about 240 measured Pension Funds. It was enough to take us from seventh to fourth place over the six months so we retained the portfolio.

It's worth reflecting on how it was done as these times will come again. As the fears of the oil crisis in 1979 subsided there was a timely large fall in yields on UK gilts (bonds). In those days few investors appreciated the sensitivity of capital values of longer dated fixed interest securities to sudden large moves in interest rates. Gilts were often lumped in with cash as an asset class.

To encourage buying, the authorities had just issued a novel 20% partly paid ten year security. I geared a high portion of the gilt portfolio into this security which promptly doubled in value prior to the second instalment. In the equity portfolio, the combination of Racal, United Scientific, Amstrad, together with two speculative high fliers in the shape of London & Liverpool and Polly Peck saved the day.

The latter two were quite dramatic. Polly Peck was trading at 7p when Asir Nadir came in for lunch. It ran to a high of £3.50 by the time I left for Australia in March 1981. Six weeks later it hit an all-time high of about £36.

In two instances I inadvertently breached investment guidelines for the maximum amount in one security. There was a prudential limit of 5% by market value in any one security. In normal circumstances a marked to market approach is fine. However, while I could control the initial outlay, I had no control over the subsequent market price. On separate occasions firstly United Scientific, then Polly Peck rose dramatically in price. I sold a few, but the price rose dramatically further between the close of the quarter and the reporting date.

Broad diversification is not so important in personal portfolios and closed end portfolios as it is with open-ended funds such as unit trusts (mutual funds). When an investor buys into a unit trust, the trust issues units. The investor is effectively buying into a second hand portfolio. Unit trusts treat all investors as equal, and issue new units. However each unitholder participates in the income and gains equally, regardless of the timing of their entry. Unit trusts are obliged to distribute all realised gains and income received each year, or otherwise pay tax at source. They invariably opt to distribute the income and realised gains. The problem is that the more successful the trust is during the year, the greater the disadvantage facing a new investor in the run-up to June 30th.

Lets say you decide to invest in a high performing trust in June. It has earned a 30% gain on its investments during the year plus a 5% dividend yield. If all these gains are realised, you will receive about 35% of your initial investment back in August as a realised distribution, even though the value of the investments may have fallen subsequently. This second hand portfolio effect is one reason why unit trusts generally hold more than 25 securities.